• Register
  • Help
Page 1 of 2 12 LastLast
Results 1 to 10 of 13

Topic: Using soundtrack db numbers: A Test

Share/Bookmark
  1. #1

    Using soundtrack db numbers: A Test

    Hi everyone -

    I\'m working on an orchestral piece using various sample libraries, and am continually trying to improve the quality to match commercial releases.

    Although I\'ve written and recorded a lot of pop material (where approaching 0 db is very popular for airplay), I am thinking it\'s not so popular for orchestral work.

    As a test, I imported track 2 from original LOTR soundtrack, which has a *volume* level / expressiveness that I am using on a new piece of mine.

    Before, I would simply use compression, or multiband compression, or an ad-limiter to push my final mix up to 0 db.

    However, track 2 of LOTR seems to hover around -9db, with peaks up around -7 and -5.8.

    Is this one of the secrets of orchestral work with samples? Should I by trying to keep my *volume* in the range where other commercial works are?? And perhaps, as a result, avoid final mix compression/mb compression/etc?

    Very interested,

    Eric Doggett
    doggett@doggettstudios.com
    www.doggettstudios.com

  2. #2

    Re: Using soundtrack db numbers: A Test

    Remember that LOTR is a soundtrack and needs to be mixed to levels that match the dialog and action in the movie. That doesn\'t necessarily translate to how you would want to adjust your own mix.

    I think a good orchestral CD would be a better choice for using as a guide.

    -- Martin

  3. #3

    Re: Using soundtrack db numbers: A Test

    In most situations, compression/limiting is a no no with orchestral and classical music. This same rule also applies to vocals in opera music as well.

    Kip McGinnis
    Bardstown Audio
    www.bardstownaudio.com

  4. #4

    Re: Using soundtrack db numbers: A Test

    Just because it came from the movie doesn\'t mean the released CD is mixed to the same level as the movie does it?

  5. #5

    Re: Using soundtrack db numbers: A Test

    you\'re right timzy.

    I\'ve read a couple of articles that they mix more dry signal in some actual movie scores, and toss more verb in on the CDs. Doesn\'t mean every movie is done this way, but it prolly varies from post mixer to post mixer.

  6. #6

    Re: Using soundtrack db numbers: A Test

    I believe most film scores will peak at 0db somewhere in the score - of course the quieter cues won\'t be normalized to 0 db - it would make a very bad listening experience if a cue dominated by p strings and a p solo flute was as loud as one with a full fff brass arrangement.

  7. #7

    Re: Using soundtrack db numbers: A Test

    I bet there\'s at least one point on the LOTR CD that it hits just under 0dB full scale. That peak might only last for one sample on the whole CD though so what it means is the CD probably isn\'t compressed, limited, loudness maximized much if at all. I\'ve been quite surprised how dynamic some recent soundtrack CDs like Harry Potter have been. I bet they\'re not that dynamic in the movie! IMO this is a good thing BTW.

    I don\'t see that the peak or average level of the CD would necessarily bear any relationship to what happens in the movie soundtrack. It is feasible that some or all of the movie soundtrack is compressed or levels ridden on the mix so as to sit with all the other audio, dialog, FX etc. The CD could be left with its original dynamics intact.

    Also, the movie soundtrack music might have been mixed using stems rather than a stereo mix, for example, a submix of the strings, one of the brass etc. so again the mix on the CD might have different dynamics from the movie sound.

    My point is that if your music is stand-alone and heard in isolation just make sure the loudest point in the music hits 0db (or just under) and don\'t use compression, loudness maxing etc. Leave the dynamics of the music alone in effect. If your music is going to be used in a game soundtrack, TV or film where it is competing with or accompanying other audio then think about compression etc. This means the quiet bits don\'t get lost and the loud peaks don\'t intrude. Hope this isn\'t old news to you!

  8. #8
    Senior Member Bruce A. Richardson's Avatar
    Join Date
    Sep 1999
    Location
    Dallas, Texas
    Posts
    5,755

    Re: Using soundtrack db numbers: A Test

    Originally posted by Eric Doggett:
    Is this one of the secrets of orchestral work with samples? Should I by trying to keep my *volume* in the range where other commercial works are?? And perhaps, as a result, avoid final mix compression/mb compression/etc?

    [/URL]
    <font size=\"2\" face=\"Verdana, Arial\">Hi Eric,

    You use the word \"avoid\" like compression is bad.

    Actually it is not. Be aware that sculpting orchestral music from samplers is a totally different playing field than recording classical music. Sampled orchestral music is electronic music, emulating acoustic, therefore, any tool in the toolbox which moves you closer to the end goal of high quality is fair game and desirable.

    Soundtracks and soundtrack albums are a very difficult genre in which to seek and find standards. There really are none. The stems for the film mix may go onto disc with nothing more than leveling. They may be totally remixed. Since film music must exist in balance with foley, dialog, nat sound, and picture, there is generally a huge amount of scoop and in some cases dynamic alterations in the vocal bands. Even in terms of orchestration, an experienced composer of scores is going to leave that space to avoid saturating the entire audio bandwidth--otherwise, the mix engineer will simply reduce the volume until all elements appear or compress the living snot out of that range.

    I don\'t do a lot of mastering work these days, just a few albums a year, but I\'ve probably done about twenty orchestral albums. I have used compression/multiband/limiting on every single one. There are just too many phenomena occurring in a digital recording to exclude these tools. One obvious for instance: Suppose there is a huge bass drum hit. Upon analysis, you discover that there is a 2 msec peak which is 12 dB louder than anything else in the track. That information is probably driving the entire track down for no musical reason at all. A 2 msec peak probably won\'t even get your woofer in motion. So you want to use some limiting, so that the information in the track is all contributing musically. Just because something exists in the track doesn\'t necessarily mean it will deliver a musical result.

    Just keep in mind that all your hard work will eventually just cause cardboard cones to vibrate in VERY low tech boxes. Puts it all in perspective. [img]images/icons/wink.gif[/img]

  9. #9

    Re: Using soundtrack db numbers: A Test

    Good points, Bruce (and everyone else!)

    I guess what I am wondering is not so much the correlation between the cd and the movie or whether or not it matches up, but rather, my own material. For example, on earlier demo cd\'s, I would attempt to get close to 0db on everything (orchestral demo pieces, rock, synth, etc). The cd would have one 4-minute or so track, mixing between styles. My first instinct was that everything on the demo had to be as loud as everything else.

    I am now wondering whether or not orchestral sampling should *not* strive to be close to 0 db at all (hence my thought on avoiding compressive/normalizing techniques).

  10. #10

    Re: Using soundtrack db numbers: A Test

    Originally posted by Eric Doggett:
    Good points, Bruce (and everyone else!)

    I guess what I am wondering is not so much the correlation between the cd and the movie or whether or not it matches up, but rather, my own material. For example, on earlier demo cd\'s, I would attempt to get close to 0db on everything (orchestral demo pieces, rock, synth, etc). The cd would have one 4-minute or so track, mixing between styles. My first instinct was that everything on the demo had to be as loud as everything else.

    I am now wondering whether or not orchestral sampling should *not* strive to be close to 0 db at all (hence my thought on avoiding compressive/normalizing techniques).
    <font size=\"2\" face=\"Verdana, Arial\">Eric
    I would say the loudest peak in your recordings should definitely hit a fraction under 0db (when you actually record or bounce the mix) whatever the style of music. It\'s just good engineering practice making full use of the available resolution.

    I would always avoid normalizing a mix. You should never need to normalize a properly recorded mix after the fact as the loudest peak should be hitting 0db anyway. Normalizing can be quite detrimental to sound quality and floating point processing of audio using a plug in like the Waves L2 for example is a better way of doing this if you really need to bring a mix up to 0dB that has been under recorded.

Go Back to forum

Tags for this Thread

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •